The Mann-Whitney U test was used to test for significant differences in continuous patient characteristics between men who received sextant TRUS-Bx (group 1) and men in the MRI side study (group 2). The χ2 test was used to test for significant differences in categorical patient characteristics and PCa detection rates between men in group 1 and 2. Cross tabu- lation was performed of 12-core TRUS-Bx (group 2a) vs MRI ± TBx (group 2b) outcomes in men in group 2 (supplementary tables 1 – 3).The McNemar’s test was used to test for differences in PCa rates between group 2a and 2b. Targeted biopsy cores of hypoechoic lesions were analyzed as part of the TRUS-Bx. The risk of a biopsy positive for (high-grade) PCa was retrospectively assessed in all men using the RPCRC4, a prediction model based on data of 2896 men who were screened for the 2nd time 4 years after initial screening in the ERSPC Rotterdam (12). Men were stratified based on the RPCRC4 risk cut-off values to perform a biopsy: a risk of any PCa ≥20% and/or a risk of high-grade (GS ≥3+4) and/ or locally advanced (T-stage ≥2C) PCa >3% (12). The potential reduction of MRI’s, biopsies and overdiagnosis of low-grade PCa by upfront risk-stratification was assessed. Statistical tests were two sided with the criterion of significance set at p <0.05. Statistical analyses were performed with SPSS for Windows (Version 21.0. Armonk, NY:IBM Corp.).
RESULTS
Patient characteristics
A total of 1734 invited men attended for PSA testing (figure 1). A total of 23% (406/1734) attendees were eligible for biopsy based on PSA ≥3.0 ng/ml and 83% (337/406) of these men were biopsied, all of whom were MRI naïve. A total of 16% (65/406) men refused a biopsy based on anticipated burden or presence of comorbidities, while the biopsy failed in 2 men due to anal sphincter tension and was omitted in 2 men in whom MRI revealed a bladder tumor. A 6-core TRUS-Bx was performed in 179 men (group 1), of whom 62 men received TRUS-Bx after initiation of the side study: 5% (10/220) men biopsied after side study initiation had a contraindication for MRI (8 with magnetic prosthetic material, 1 with a pacemaker and 1 with GFR <30 mL/min) while 24% (52/220) men were not willing to undergo MRI, most frequently due to the extra (MRI-)visit additional to the biopsy visit and claustrophobia. A total of 158 men received MRI and 12-core TRUS-Bx ± TBx (group 2). The median age and PSA were 73 years (IQR 72 – 74) and 4.3 ng/ml (IQR 3.4 – 5.7) respectively (table 1). A total of 55% (184/337) men had received ≥1 TRUS-Bx at previous screening rounds. There were no significant differences in patient characteristics between men in group 1 and 2.
56
Prostate cancer detection at fifth screening (primary outcome)
A total of 36% (120/337) biopsied men were diagnosed with PCa. The low-grade PCa rate was 23% (77/337) and the high-grade PCa rate was 13% (43/337) (table 2). In biopsy naïve men the overall, low-grade and high-grade PCa rates were 48% (74/153), 27% (41/153) and 22% (33/153), respectively. The overall, low-grade and high-grade PCa rates in previ- ously biopsied men were 25% (46/184), 20% (36/184) and 5% (10/184), respectively.
Prostate cancer detection of different biopsy strategies (secondary outcome)
The median number of biopsy cores in group 1, 2, 2a and 2b was respectively 6 (IQR 6 – 6), 12 (IQR 12 – 14), 12 (IQR 12 – 12) and 2 (IQR 2 – 3). A total of 30% (48/158) men in group 2 had a positive MRI and received MRI-TBx (table 3). The high-grade PCa rate was not sig- nificantly different between group 1 and 2 (11% vs 15%, p = 0.21). Sextant TRUS-Bx (group 1), 12-core TRUS-Bx (group 2a) and MRI ± TBx (group 2b) had a similar high-grade PCa rate of respectively 11%, 12% and 11% (table 3). The low-grade PCa rate was significantly
5th screening round attendees 1734
PSA < 3.0 ng/ml
1328 PSA ≥ 3.0 ng/ml 406
No biopsy
60 Sextant TRUS-Bx 179 12-core TRUS-Bx ± TBx 158 Multiparametric MRI
167
No biopsy 9 Sextant TRUS-Bx outcomes
• No PCa: 130 (73%) • GS 3+3 PCa: 30 (17%) • GS ≥3+4 PCa: 19 (11%)
12-core TRUS-Bx outcomes • No PCa: 94 (59%) • GS 3+3 PCa: 45 (28%) • GS ≥3+4 PCa: 19 (12%) TBx outcomes • No TBx: 110 (70%) • No PCa: 20 (13%) • GS 3+3 PCa: 11 (7%) • GS ≥3+4 PCa: 17 (11%) 12-core TRUS-Bx ± TBx outcomes • No PCa: 87 (55%)
• GS 3+3 PCa: 47 (30%) • GS ≥3+4 PCa: 24 (15%)
Figure 1. Flowchart of all attendees in the 5th screening round of the ERSPC Rotterdam PSA = prostate-
specific antigen, MRI = Magnetic Resonance Imaging, TRUS-Bx = transrectal ultrasound-guided systematic bi- opsy, TBx = targeted biopsy, PCa = prostate cancer, GS = Gleason score
57
3
Table 2. Prostate cancer detection rates at 5th screening stratified for previous biopsy status
Prostate cancer detection rates at 5th screening in the ERSPC Rotterdam
Sextant TRUS-Bx (Group 1)
12-core TRUS-Bx with MRI ± TBx (Group 2)
All men (Group 1 + 2)
Number % Number % p-value Number %
biopsy naïve men
GS 3+3 PCa GS ≥3+4 PCa All 18 14 79 23 18 100 23 19 74 31 26 100 0.25 0.23 41 33 153 27 22 100
Previously biopsied men
GS 3+3 PCa GS ≥3+4 PCa All 12 5 100 12 5 100 24 5 84 29 6 100 <0.01 0.78 36 10 184 20 5 100
biopsy naïve + previously biopsied men GS 3+3 PCa GS ≥3+4 PCa All 30 19 179 17 11 100 47 24 158 30 15 100 <0.01 0.21 77 43 337 23 13 100 ERSPC = European Randomized study of Screening for Prostate Cancer, TRUS-Bx = transrectal ultrasound-guid- ed systematic biopsy, MRI-TBx = MRI-targeted biopsy, GS = Gleason score, PCa = prostate cancer
Table 1. Patient characteristics of all men who received a biopsy in the 5th screening round of the ERSPC
Rotterdam
Patient characteristics of all biopsied men in the 5th screening round
Sextant TRUS-Bx (Group 1) 12-core TRUS-Bx ± MRI-TBx (Group 2) All men
Median IQR Median IQR p-value Median IQR
Age (years) 73.2 72.2 – 74.1 73.1 72.4 – 73.8 0.82 73.1 72.3 – 74.0 PSA (ng/ml) 4.4 3.5 – 5.6 4.2 3.4 – 5.8 0.96 4.3 3.4 – 5.7 Prostate volume (ml) 48.3 35.3 – 68.4 52.9 36.8 – 70.9 0.53 50.2 35.9 – 70.1 Number of previous TRUS-Bx 1 0 – 2 1 0 – 2 0.48 1 0 – 2
Number % Number % p-value Number %
T-stage DRE T0 158 88 134 85 0.35 292 87 T2a 15 8 9 6 0.34 24 7 ≥T2b 6 3 15 9 0.02 21 6 T-stage TRUS T0 151 84 135 85 0.78 286 85 T2a 25 14 20 13 0.73 45 13 ≥T2b 3 2 3 2 0.88 6 2 Previous TRUS-Bx 0 79 44 74 47 0.62 153 45 1 48 27 35 22 0.32 83 25 2 29 16 34 22 0.21 63 19 3 15 8 12 8 0.79 27 8 4 8 4 3 2 0.19 11 3 Total 179 100 158 100 337 100
TRUS = transrectal ultrasound, TRUS-Bx = transrectal ultrasound-guided systematic biopsy, MRI-TBx = MRI-tar- geted biopsy, IQR = Interquartile range, PSA = prostate-specific antigen, DRE = digital rectal examination
58
higher in group 2 compared with group 1 (30% vs 17%, p <0.01). The low-grade PCa rate of 12-core TRUS-Bx was significantly higher compared with sextant TRUS-Bx (28% vs 17%, p = 0.01), whereas the low-grade PCa rate of MRI ± TBx was significantly lower (7%, p <0.01).
Table 3. Prostate cancer detection rates using different biopsy strategies without upfront risk-stratification
PCa detection rates using different biopsy strategies without upfront multivariable risk-stratification
Sextant TRUS-Bx (Group 1; n=179)
12-core TRUS-Bx with MRI ± TBx (Group 2; n=158) 12-core TRUS-Bx (Group 2a; n=158) MRI ± TBx (Group 2b; n=158) number % number % number % number %
No biopsy - - - 110 70
No PCa 130 73 87 55 94 59 20 13
GS 3+3 PCa 30 17 47 30 45 28 11 7 GS ≥ 3+4 PCa 19 11 24 15 19 12 17 11
Total 179 100 158 100 158 100 158 100 TRUS-Bx = transrectal ultrasound-guided systematic biopsy, MRI = Magnetic Resonance Imaging, TBx = targeted biopsy, PCa = prostate cancer, GS = Gleason score
TRUS-bx versus MRI ± Tbx outcomes in group 2
A total of 15% (24/158) men in group 2 were diagnosed with high-grade PCa. High-grade PCa was detected by TRUS-Bx only, MRI ± TBx only, or both in respectively 29% (7/24), 21% (5/24) and 50% (12/24) (supplementary table 1). An MRI ± TBx only and TRUS-Bx only strat- egy would have missed high-grade PCa in respectively 4% (7/158) and 3% (5/158) men. A total of 30% (47/158) men in group 2 were diagnosed with low-grade PCa. Low-grade PCa was detected by TRUS-Bx only, MRI ± TBx only, or both in respectively 77% (36/47), 11% (5/47) and 13% (6/47).
Avoidance of MRI’s, biopsies and overdiagnosis by upfront risk-stratification (tertiary outcome)
Upfront RPCRC4-based stratification in group 1 would have avoided 67% (120/179) biop- sies, while missing 33% (10/30) low-grade PCa and 26% (5/19) of high-grade PCa (table 4). RPCRC4-based stratification in group 2 would have avoided 65% (102/158) MRI’s and bi- opsies, missing 68% (32/47) low-grade PCa and 17% (4/24) high-grade PCa. RPCRC4-based stratification in group 2a would have avoided an equal percentage of biopsies (65%) and low-grade PCa (66%), missing 29% (7/24) high-grade PCa. Although RPCRC4-based strati- fication in group 2b would have missed 42% (10/24) high-grade PCa, almost all low-grade PCa diagnoses (94%; 44/47) would have been avoided.
59
3
biopsy complications
A total of 5.6% (19/337) men had post-biopsy fever, of whom 3.3% (11/337) were admit- ted. The urosepsis admission rate was 3.4% (6/179) in group 1 and 3.2% (5/158) in group 2. One man was admitted for urinary retention (group 1) and another for a transient ischemic attack one day post-biopsy (group 2). No deaths occurred.
dISCUSSIOn
The benefit of a mortality reduction achieved by screening must be balanced against the harms of unnecessary biopsies and overdiagnosis, particularly in elderly men (7, 8). Although the American Urological Association recommends against routine screening in men aged ≥70 years (24), older men are frequently screened in daily practice (25, 26). The European Association of Urology recommends an individualized risk-adapted strategy for early detection in men with a good performance status and life expectancy ≥10 – 15 years (27). As the median life expectancy still increases and the incidence of poorly-differenti- ated PCa increases with age, some men aged ≥70 years may still benefit from screening (9-11). Hence, patient-selection before biopsy is mandatory in this age group and could be achieved by the use of risk-stratification and MRI. In the present study we explored the potential such a strategy in repeatedly screened men aged 71 – 75 years in the ERSPC Rotterdam 5th screening round.
Table 4. Percentages of avoided MRI’s, biopsies and low-grade PCa by multivariable risk-stratification with
the Rotterdam Prostate Cancer Risk Calculator 4.
Avoided MRI’s, biopsy procedures and overdiagnosis of low-grade PCa by multivariable risk- stratification RPCRC4 + Sextant TRUS-Bx (Group 1; n=179) RPCRC4 + 12-core TRUS-Bx with MRI ± TBx (Group 2; n=158) RPCRC4 + 12-core TRUS-Bx (Group 2a; n=158) RPCRC4 + MRI ± TBx (Group 2b; n=158)
number number number % number % number % MRI avoided - - 102/158 65 - - 102/158 65 Biopsy procedure avoided 120/179 67 102/158 65 102/158 65 130/158 82 GS 3+3 PCa diagnosis avoided
Biopsy strategy = ref. 12-core TRUS-Bx ± TBx = ref.
10/30 - 33 - 32/47 32/47 68 68 29/45 31/47 64 66 8/11 44/47 73 94 GS ≥3+4 PCa diagnosis missed
Biopsy strategy = ref. 12-core TRUS-Bx ± TBx = ref.
5/19 - 26 - 4/24 4/24 17 17 2/19 7/24 11 29 3/17 10/24 18 42 RPCRC4 = Rotterdam Prostate Cancer Risk Calculator 4, TRUS-Bx = transrectal ultrasound-guided systematic biopsy, MRI = Magnetic Resonance Imaging, TBx = targeted biopsy, PCa = prostate cancer, GS = Gleason score, ref. = reference.
60
The overall PPV in the 5th screening round was 36% (120/337) and thus considerably higher than the PPV in the first (25.5%) and consecutive (≈20%) screening rounds of the ERSPC Rotterdam (3). The high PPV at 5th screening could be explained by the fact that half of men received 12-core TRUS-Bx ± MRI-TBx (PPV = 45%; 71/158) instead of 6-core TRUS-Bx (PPV = 27%; 49/179). The 27% overall PPV of 6-core TRUS-Bx is still higher than expected based on previous screening rounds, probably due to the higher age at 5th screening (9-11). The overall PPV in biopsy naïve men was ≈50% and approximately half of PCa detected in these men were high-grade. In line with findings at previous screening rounds, it is evident that biopsy naïve men after attending 4 rounds of screening with a 4-year interval (i.e. predominantly men with PSA <3.0 ng/ml for a long period) still can be diagnosed with high-grade disease (3). In addition, the PPV in previously biopsied men was still considerable (25%), most likely reflecting the known undersampling with 6-core TRUS-Bx (6). In line with previous screening rounds, 20% of PCa detected in previously biopsied men were high-grade (3). However, the fact that only 5% of previously biopsied men were actually diagnosed with high-grade PCa questions whether these men should be repeatedly screened using the PSA-based random TRUS-Bx algorithm.
Half of men in the 5th screening round received sextant TRUS-Bx and half received 12-core TRUS-Bx and MRI ± TBx, allowing a comparison of different biopsy strategies. In a clinical setting, it was shown that 6-core TRUS-Bx undersamples (high-grade) PCa as compared with 12-core TRUS-Bx (6). Furthermore, MRI ± TBx detects significantly less low-grade PCa than TRUS-Bx (18, 19). The high-grade PCa detection of MRI ± TBx is at least equal com- pared with TRUS-Bx in a clinical setting (18, 19). Recently, a small pilot study in the 10th screening round of the Göteborg trial investigated the value of MRI in a population-based screening setting (28). In men with PSA ≥3.0 ng/ml the low-grade PCa rates of 10-core TRUS-Bx and MRI ± TBx (cognitive) were respectively 13% (9/70) and 8% (5/65), while the high-grade PCa rates were respectively 13% (9/70) and 11% (7/65) (28). Accordingly, a similar high-grade PCa rate of sextant TRUS-Bx, 12-core TRUS-Bx and MRI ± TBx(fusion) was found in the present study (11%, 12% and 11%). Surprisingly, only the low-grade PCa rate of 12-core TRUS-Bx was higher compared with 6-core TRUS-Bx (28% vs 17%, p =0.01), suggesting that the use of an extended TRUS-Bx scheme would only increase the detection of low-grade disease in previously screened (and often biopsied) men. Less surprisingly, the low-grade PCa rate of MRI ± TBx was significantly lower (7%) as compared with 6-core TRUS-Bx (17%) and 12-core TRUS-Bx (28%), indicating that a screening strategy with only MRI ± TBx could significantly reduce overdiagnosis of low-grade disease. Furthermore, an MRI ± TBx strategy would cause a significant reduction of biopsies as more than two-thirds of population-based screened men with PSA ≥3.0 ng/ml have a negative MRI (28).
61